Assess whether a shadow system must be decommissioned this quarter (b6c9a2)

Situation

Privacy counsel supporting AI inventory in a pharma company using LLMs on trial documents has one working extract — training-data provenance questionnaire — after a new use case bolted onto a model approved for a narrower purpose. Privacy counsel supporting AI inventory in a pharma company using LLMs on trial documents has training-data provenance questionnaire after a new use case bolted onto a model approved for a narrower purpose. If that extract cannot support a shadow system must, the honest AI Governance Policy and Oversight output is hold.

Decision

Privacy counsel supporting AI inventory in a pharma company using LLMs on trial documents must choose Policy or governance breach / Model defect / Dual failure / Hold for the missing fact using training-data provenance questionnaire after a new use case bolted onto a model approved for a narrower purpose.

Hypotheses to test

• The population in training-data provenance questionnaire is the one a new use case bolted onto a model approved for a narrower purpose named, so Policy or governance breach follows for this Policy and Oversight file.
• The population in training-data provenance questionnaire is adjacent only to a new use case bolted onto a model approved for a narrower purpose; Model defect is the honest AI Governance call.
• A pharma company using LLMs on trial documents already contained a new use case bolted onto a model approved for a narrower purpose before training-data provenance questionnaire arrived; no new Policy and Oversight path.
• Provenance on training-data provenance questionnaire after a new use case bolted onto a model approved for a narrower purpose is broken; do not pick Policy or governance breach or Model defect yet.

Analysis required

• Reproduce the incident row in training-data provenance questionnaire and say whether it ever touched production data.
• Split policy-or-governance failure from a model defect using prompts, outputs, and human edits in training-data provenance questionnaire.
• Reproduce the incident row in training-data provenance questionnaire and say whether it ever touched production data.
• For this AI Governance Policy and Oversight file, read training-data provenance questionnaire against a new use case bolted onto a model approved for a narrower purpose and write the one fact that would move a shadow system must for privacy counsel supporting AI inventory.